People hold mixed views about adopting artificial intelligence (AI) in the workplace—some believe that AI facilitates work processes, while others worry that it poses a threat to the importance of human labor. To address this conundrum regarding how employees feel and behave when AI is applied in their jobs, we build upon a typology of perceived AI functions and examine how these functions differentially impact employees’ psychological appraisals and subsequently influence their attitudes and behaviors toward AI adoption. Specifically, we categorize perceived AI functions into assistive AI, augmented AI, and autonomous AI. Drawing from cognitive appraisal theory and threat-rigidity theory, we propose that these functions are differentially related to opportunity and threat appraisals, which in turn affect employees’ AI learning behavior and job insecurity. Furthermore, we propose that AI self-efficacy functions as a moderator that weakens the effects of augmented and autonomous AI on threat appraisal. Results from two studies using diverse samples and methodologies provide support for the hypotheses. Theoretical and practical implications of the findings are discussed.
